Introduction

The glymphatic system is a paravascular cerebrospinal fluid (CSF) pathway that drives waste clearance in the brain. Glymphatic dysfunction has been associated with various neurodegenerative diseases,¹ highlighting the importance of evaluating the glymphatic function in vivo. Dynamic contrast-enhanced MRI (DCE-MRI) provides the opportunity to assess whole-brain glymphatic dynamics in vivo by directly tracking the transport of CSF tracers.² A previous study in rats showed significant differences in the influx and transport of two Gd-based contrast agents with different molecular weights (MW), Gd-DTPA (MW=661.8 Da) and GadoSpin (MW=200 kDa).³ A recent study using ¹⁷O-enriched water (H₂¹⁷O, MW=19 Da) reported strikingly increased transport of H₂¹⁷O over Gd-DTPA in the rat brain.⁴ These studies suggest that delineation of the glymphatic dynamics is dependent on the molecular size of the CSF tracer.
In the current study, we compared the transport kinetics of intracisternally administered Gd-DTPA, GadoSpin, and H₂¹⁷O in the mouse brain. Our results showed drastically different kinetics of the three tracers in both the brain parenchyma and lateral ventricle, suggesting size-dependent pathways and kinetics of molecular transport in the glymphatic system.

Methods

Experimental Setup: All MRI data were acquired on a Bruker 9.4T scanner. Adult C57BL/6 mice were randomly assigned to three groups with intracisternal infusion of 1) 12.5 mM Gd-DTPA (n=2); 2) 6.25 mM GadoSpin (n=2); and 3) 40% H₂¹⁷O (n=2). Mice were anesthetized with isoflurane. A polyethylene catheter was inserted into the cisterna magna and secured with cyanoacrylate glue.⁵ The animal was then transferred to the scanner. Following the baseline scans, 10 μL of contrast agent was infused at 1 μL/min. The transport of the contrast agents in the brain was assessed using either T₁-weighted (Gd-DTPA and GadoSpin) or T₂-weighted (H₂¹⁷O) MRI. To validate the intracisternal injection method, two additional mice were injected with either 5% Evans blue or a cocktail of FITC-dextran (MW=2000 kDa) and CF594 hydrazide (MW=759 Da) and imaged with bright-field microscopy and fluorescent cryoimaging, respectively.
Gd-DTPA and GadoSpin Imaging: Gd-DTPA and GadoSpin were tracked using a T₁-weighted 3D FLASH sequence with the following parameters: TR/TE, 50/2.5 ms; flip angle, 15°; FOV: 20×16×14 mm³; matrix, 200x80x70; NAV, 1, leading to an isotropic resolution of 200 μm and a temporal resolution of <5 min. 26 images were acquired continuously for ~130 min.
H₂¹⁷O imaging: H₂¹⁷O transport was imaged with a T₂-weighted multislice RARE sequence. 28 slices were acquired with TR, 2500 ms; FOV, 20x16 mm²; matrix, 200x80; slice thickness, 0.5 mm. 8 echoes were acquired in each TR with an effective TE of 31 ms, leading to 25-s acquisition time. A total of 312 datasets were acquired continuously for ~130 min.
Data Analysis: All data analyses were performed using in-house developed MATLAB code. Baseline-normalized signal changes over time and the percentage signal changes from the baseline were evaluated in selected ROIs including the olfactory bulbs (OB), the lateral ventricle (LV), and the perfusion territory of the middle cerebral artery (MCA).

Results

Intracisternal administration of CSF tracers was confirmed by microscopy and cryoimaging (Figure 1). Evan blue and FITC-dextran were confined to the CSF space, while CF594 was transported into the parenchyma. Figure 2 shows representative images from mice injected with Gd-DTPA, GadoSpin, and H₂¹⁷O, respectively. Shown in Figure 3 are the time maximum intensity projection (tMIP) for Gd-DTPA and GadoSpin, and the time maximum intensity change projection (tMICP) for H₂¹⁷O. Gd-DTPA was first observed in the CSF-filled aqueduct, and then progressed along the ventral side of the brain in the subarachnoid space. Transport of Gd-DTPA into the parenchyma can be appreciated after 20 min. In contrast, GadoSpin was confined to the CSF spaces, while H₂¹⁷O showed more extensive penetration into the parenchyma.
Figure 4 shows the time courses of signal changes in selected ROIs. Gd-DTPA and GadoSpin induced a signal increase in T₁-weighted images, while H₂¹⁷O caused a decrease in T₂-weighted images (Figure 4A). The time courses of percentage signal change from the baseline reflect the kinetics of contrast agent transport (Figure 4B). GadoSpin transport to the parenchyma (OB and MCA territory) was minimal, while the uptake of Gd-DTPA and H₂¹⁷O was delayed in the distal parenchyma (OB). Compared to Gd-DTPA, H₂¹⁷O showed a much faster uptake rate in both OB and MCA territory. Interestingly, the three tracers showed drastically different transport kinetics in LV. H₂¹⁷O transport to LV reached equilibrium rapidly, while Gd-DTPA showed continued accumulation over the entire course of imaging. Further, GadoSpin transport to LV was the fastest, accompanied by rapid washout. These observations suggest different transport pathways and kinetics in the glymphatic system that is size-dependent.

Discussion and Conclusion

Consistent with previous observations in rats^3,4,6, GadoSpin transport in mice was confined to the CSF space, and H₂¹⁷O uptake in the parenchyma was faster than Gd-DTPA. In addition, remarkably different transport kinetics of the three contrast agents to LV were observed, suggesting the existence of size-dependent pathways for molecular transport in the glymphatic system. Our current work lays the foundation for future studies employing genetically manipulated mouse models for a mechanistic understanding of the regulation of glymphatic function in normal and diseased conditions.